Agentic commerce: how to maintain control over a purchase journey that brands no longer control?

By Camille Daudet, Senior Manager AI and Transformation at Converteo

We were promised that AI would revolutionize purchase journeys. And the signs are already here:

• by 2030, AI agents could manage up to $5 trillion of global commerce according to McKinsey’s projections;
• already 22% of French people say they are ready to delegate their purchases to them;
• partnerships are multiplying (Expedia, Le Bon Coin, Accor, Carrefour…);
• and the technology has reached a sufficient level of maturity to intervene at any stage of the consumer journey – from gift inspiration to the purchase itself.

Yet when it comes to organizing a trip, choosing a new car, or buying a birthday present, something resists. A consumer reflex: do I really want to delegate that?

It is precisely this ambivalence that is at the heart of agentic commerce today, and it is the consumer who decides the role that the brand and the AI agents play in their purchase journey.

AI Agents & Brands: The Risk of Disintermediation

By capturing the consumer’s intent earlier and earlier in their journey – a birthday, a travel plan… – agents are indeed positioning themselves as key players in a brand’s visibility.

Thus, a ChatGPT user wishing to organize their next trip is now suggested the services of Expedia as a priority, without even having mentioned it in their search.

To emerge as a brand in such an environment, it will be necessary to address the agents as consumers in their own right and learn to speak their language, without neglecting the quality of one’s offer.

To delegate or not to delegate to agents: that is the question

This risk of disintermediation must, however, be nuanced. The consumer, of course, remains in control of their journey, considering the AI agent as a new purchasing channel at their disposal to which they can delegate certain tasks.

Several criteria come into play in the decision to delegate or not to the agent: the tediousness of the task for recurring purchases, the time available, but also the pleasure and emotion linked to the purchase in question.

The same consumer might unhesitatingly delegate the organization of their recurring business trips to an AI but will keep control over all or part of the preparation for their honeymoon.

The challenge for brands is therefore not to consider the agent as the sole entry point, but on the one hand to identify where and why it creates value for its consumers: for which purchase, for which customer profile, and at which stage of the journey? And, on the other hand, to create a flexible purchase journey that gives them complete freedom to choose their level of intimacy with the brand and the degree of intervention of AI agents.

6 shades of agentic to adapt to consumer expectations

Moreover, AI agents have all the technical capabilities to meet consumer expectations, as McKinsey recalls with the concept of the “agentic commerce automation curve“.

• At the lowest level of the curve, that of minimal delegation: the agent is able to manage pre-programmed recurring purchases, like monthly grocery shopping.
• At the other extreme of the curve, it is human intervention that becomes minimal: several agents are thus mandated by consumers to negotiate among themselves and carry out complex purchases such as subscribing to an insurance contract.
• Between the two, four levels that match the degree of autonomy desired by the consumer in each situation.

It is precisely this granularity that makes agentic commerce viable: the agent does not impose itself, it adapts.

The AI agent, a consumer in its own right?

Commerce must not, of course, be only agentic, and brands will have to rethink the purchase journey by combining a direct relationship with consumers and with AI agents.

More than an additional marketing channel, the AI agent is becoming a new kind of consumer: pragmatic, autonomous, and guided by the clarity of information and efficiency.

To respond to this, brands and distributors will have to simplify access to their data and sustainably rethink their technical architecture and organization, in order to allow agents to interact wisely with their ecosystem – and for consumers to navigate freely between the website, agents, and sales advisors.